Observing Summary: Target RA Dec K AB Conguration/mode Days 180 NEARBY SINGLE OTHER 45 STARS, D<10PC 500 NEARBY SINGLE OTHER 21 STARS, D<20PC 70 BROWN

Transcription

1 Next Generation Space Telescope Ad-Hoc Science Working Group Design Reference Mission Proposal Program contacts: John Trauger, Raghvendra Sahai, Karl Stapelfeldt, Dwight Moody, Jonathan Lunine Scientic category: EXTRASOLAR PLANETS Instruments: NIR{MIR/CAM+CORONAGRAPH Days of observation: 97 Abstract We describe a key NGST Origins science program, the study of giant planet and brown dwarf companions to single stars within 20 parsecs of the Sun. These substellar objects can be directly detected from their superthermal emission in the = 4:6, 5:1 m molecular opacity window through the use of a graded-mask Lyot coronagraph and the nominal NGST wavefront. Combining the emergent uxes from recent giant planet/brown dwarf atmosphere models, stellar data from the Gliese catalog, and a detailed model of coronagraph performance for the case of a seven segment primary mirror, we nd that NGST will be capable of imaging planetary companions of Jupiter's mass, age, and orbital semi-major axis around all single stars within 8 pc of the Sun in integration times of three hours or less. We propose (1) A survey of the nearest 180 single stars for Jupiter-like companions complete to 8 pc, (2) A more extensive survey of the nearest 500 single stars for more luminous (younger and/or more massive) companions that will be nearly complete to a 5 m ux level 4 magnitudes brighter than Jupiter, and (3) Detailed spectrophotometric study of selected objects discovered in these two surveys. These observations can be expected to provide the rst determination of the giant planet/brown dwarf luminosity function, the rst spectral characterization of these objects across a broad range of eective temperatures, and the rst direct images of planets orbiting another star. The results will be of fundamental importance to our understanding of planetary systems and their frequency in the galaxy.

2 Observing Summary: Target RA Dec K AB Conguration/mode Days 180 NEARBY SINGLE OTHER 45 STARS, D<10PC 500 NEARBY SINGLE OTHER 21 STARS, D<20PC 70 BROWN OTHER 31 DWARFS/PLANETS COMPANIONS Grand total days 97 2

3 Scientic Objectives 1 Background: Brown Dwarfs and Extra-Solar Giant Planets The discovery of the cool brown dwarf GL229B (Nakajima et al. 1995) has led to a major resurgence of interest in the faint end of the stellar luminosity function. Brown dwarfs (BDs) are gaseous objects intermediate in mass between Jovian planets and stars; with masses <0.08 M, they never obtained the core pressures and temperatures required for hydrogen fusion. In addition, the exciting discoveries of planetary companions from radial velocity surveys of solar-type stars (Mayor & Queloz 1995, Marcy & Butler 1998, Cochran et al. 1997) has proved for the rst time that other planetary systems exist, and that they can be very dierent from our own. The large eccentricities and small separations found for many extrasolar giant planets (EGPs) has spurred signicant new theoretical research on the formation and orbital stability of planetary systems, and the possible similarities and dierences in the formation processes of EGPs and BDs. The detection and characterization of these substellar objects represent a new astrophysical frontier that should produce insights into several key issues such as binary star formation, accretion processes in protoplanetary disks, and the nature of dark matter. It is generally believed that BDs (like low-mass stars) form through the collapse of dense molecular cloud cores, and that planets form in circumstellar disks through coagulation of planetesimals; however we currently do not know the range of companion masses that can be produced by these two formation mechanisms, or if those two ranges overlap. An upper limit to the mass of planets may be determined by the limited supply of material available in proto-planetary disks. Unfortunately, our understanding of these issues is extremely poor because the number of detected BDs/EGPs remains quite small. A decade of extensive observational eort has been rewarded with today's capability to detect and study isolated brown dwarfs in their own instrinsic thermal emission. However, the detected objects are either relatively young (age < 1 Gyr) or massive ( few 10 M J ). For fainter (older or less massive) objects, the detection statistics are dismal: only one cold (T ef f <1000K) \methane" brown dwarf has been found so far. No BD/EGP companions have been directly detected in the AU range of separations corresponding to planets in our solar system. Signicant advances will require the ability to obtain a reasonably complete census of BDs/EGPs, both in isolation and in orbit around stars (e.g. covering a range of masses, primary star spectral type, and orbital separations). As a followup to such a survey, spectra of a range of dierent BDs/EGPs are highly desirable to constrain their physical properties, including temperature and luminosity as a function of age, their gravity, and the presence of molecules in their atmospheres. Detection and characterization of EGPs is an important rst step that NGST can provide in the study of other planetary systems { a major theme of NASA's Origins program. 3

4 2 The role of NGST Full-sky ground-based infrared surveys such as DENIS and 2MASS are steadily building up the detection statistics for isolated brown dwarfs. Large telescopes and interferometers with adaptive optics should soon make initial direct detections of BD companions of nearby stars. However, such observations will be limited to the bright end (high mass/low age) of the BD luminosity function, and to wide binary separations for the detected companions. Indirect search techniques such as radial velocity or astrometric measurements of the primary star provide important complementary information (the companion's minimum mass, and its orbital period), but also suer selection eects toward higher masses and short orbital periods. The indirect techniques provide no photons from the BDs/EGPs, leaving us to speculate as to the luminosity and spectral character of these new worlds. A major opportunity for observational progress is oered by the NGST. Its unsurpassed sensitivity at near- and mid- infrared wavelengths, combined with its high spatial resolution and stable PSF, will facilitate pioneering work that includes the detection and characterization of substellar companions of nearby stars. These capabilities are not contemplated for any other observatory that will be available before the end of the next decade. SIRTF will have the sensitivity to detect isolated BDs, but lacks the spatial resolution to distinguish them as companions on solar-system scales. Large ground-based telescopes with adaptive optics will oer the angular resolution needed to resolve BD companions from their parent stars, but their high backgrounds and contrast limitations will restrict their contributions to the most luminous objects at the largest separations. With the proper instrumentation, NGST can bridge this gap by facilitating the direct detection of substellar companions of nearby stars, including objects whose masses and ages are comparable to our own Jupiter and Saturn, and whose orbital positions fall in the range of 5-10 AU. Direct detection has the added advantage that it is no more dicult to detect a complex planetary systems with multiple planets than a single planetary companion. It provides the means to further investigate many planetary systems discovered in groundbased radial veleocity searches, including the recent discovery of three EPGs around Andromedae (Butler et al. 1999) and others that will inevitably be discovered in the next several years. In designing our observing program, we lean heavily on coronagraphic instrument performance models being developed by the NGST/ISIM HCOSS concept study group (Trauger et al. 1998). We have computed detailed estimates of the NGST optical performance that include with a measure of conservatism the optical characteristics and operational precedures presently anticipated for NGST. A seven-hexagon deployed primary mirror is assumed, following the system architecture study of TRW. Provision is made for wavefront correction on the primary mirror, with up to 394 actuators working on each hexagonal segment according to the Arizona mirror design concept (Angel and Burge 1999). The optical alignment procedures of D. Redding (private communication) are implemented as follows. First, the seven individual mirror segments are assigned surface gure errors with power spectral densities (PSDs) ten times worse than expectations based on recent experience with large astronomical mirrors. Next, the seven randomly misaligned mirror segments are brought into alignment 4

5 by wavefront retrieval of a star image. Finally, the mirror actuators are used to further rene the optical wavefront through phase retrieval of a star image, but with actuator positional errors of 10 nm added randomly to simulate the performance of the Angel and Burge (1999) actuator concept. This alignment procedure terminates once the diraction limit is achieved at 2 m wavelength. We emphasize that no further wavefront corrections are assumed for this proposal beyond this basic NGST optical requirement. 3 Proposed programs We propose an observing program to identify and characterize jovian planet and brown dwarf companions around nearby stars using an NGST coronagraph. By detecting planetary photons directly, NGST will provide the rst opportunity to spectrally characterize exoplanet atmospheres. In conjunction with mass determinations for the companions from astrometric surveys, NGST observations will allow the theoretical cooling curves for substellar objects to be compared with actual luminosity and temperature measurements. Finally, by taking the rst direct images of planets orbiting other stars, NGST will help bring NASA's Origins program to the attention of the general public. The eects of scattering and diraction in the instrument are combined with backgrounds from solar system zodiacal emission and emission from the telescope primary to estimate the background level against which a planet must be detected. Shot noise on this background and detector noise are combined together to estimate integration times for planet detection at S=N = 10. This is a somewhat conservative detection criterion, since planets will always appear as point-like images whereas residual speckles take on a spectrally dispersed and streaked appearance in broad-band images. Additional image processing with spatial ltering should provide better sensitivity, but is not assumed in our feasibility criterion. Exposure times are calculated for actual Gliese catalog objects, using their apparent mangitudes and distances. Early in the HCOSS study, it became clear that 5 m is an especially attractive wavelength for the study of Jovian planets and brown dwarfs. At this wavelength the atmospheres are signicantly cleared of molecular opacity, allowing thermal emission from warmer, deeper levels of the objects to escape directly to space. It has long been known that Jupiter's disk shows broad \hot spots" in 5 m images (Ortiz et al 1998). The spectrum of GL229B also shows a prominent ux enhancement at 5 m (Oppenheimer et al. 1998). Theoretical spectra from model brown dwarf atmospheres consistently show the same bright emission near 5 m, and indicate that these objects can be many orders of magnitude brighter than an equivalent blackbody radiator at the planet's eective temperature (Burrows et al. 1997; Allard et al. 1997). For example, the 5 m contrast between a Jupiter-like companion and an M star primary (such as Lalande 21185) is just The superthermal emission of EGPs and BDs at 5 m oers a distinct opportunity for planet detection that NGST should exploit. 5

6 (1) Search for Planets We will search for giant planet companions in a sample of nearby single stars (d<10 pc). This search will be aimed at detecting EGPs at least as bright as Jupiter (i.e. mass =0:001 M, age =4:5 Gyr, absolute M magnitude 23) at separations of 5 AU. Using a 10% bandwidth M lter, the survey will be able to detect Jupiters around all single stars within 8 pc, and around a signicant fraction of others stars at greater distances. Detection of more luminous (more massive, or younger) or more widely separated companions will be correspondingly easier, and can be accomplished for progressively more distant systems from the Sun. To observe 180 targets (90 of them within 8 pc), this program will require a total of 45 days of integration time. With integration times of 3 hr/object, the search will include detections of planets with Saturn's brightness and orbital radius in 10% of the systems within 8 pc. We will also carry out \target-of-opportunity" observations towards stars which show evidence for well-separated EGP companions, such as the those found orbiting Andromedae by Butler et al. (1999). (2) SNAPSHOT Search for Extra-Solar Giant Planets/ Brown Dwarfs We will carry out a search for brown dwarf companions in a sample of nearby single stars (d<20 pc) in the M band. This search will be aimed at detecting BDs/EGPs brighter than an absolute M magnitude of 19. For reference, either a young 1 M J object at 1 Gyr or a more massive 5 M J object at 5 Gyr would have an absolute M magnitude of 19. Using relatively short integration times (0.5 hr/object) will allow us to carry out this survey in a SNAPSHOT observing mode reminiscent of HST, for a total of 21 days of integration time. We expect to obtain a nearly complete census of BDs/EGPs ( >77%) in our sample (about 500 objects assuming a 50% attrition due to objects with stellar binary companions, and excluding those included in program 1). Since the brightnesses of BDs/EGPs depend on both their masses and their ages, each observed magnitude denes a curve in the mass-age plane. (3) Characterization of Planets and Brown Dwarfs For the brightest objects uncovered in the two surveys above, we propose spectroscopic or spectrophotometric observations from = 1, 20m. For the 50 brightest brown dwarfs, we will take R = 100 slit spectra using the coronagraph's spectral mode. For the 20 brightest EGP companions, which will be signicantly fainter, we will employ lter photometry only. The power of such observations for developing a detailed understanding of BDs has been demonstrated by recent observational and theoretical results on GL229B, where modeling of its strong methane (CH 4 ) and water (H 2 O) absorption bands between 1, 4m have provided constraints on the temperature, composition, mass, and age of the brown dwarf though comparison with a grid of atmosphere models coupled to an evolution/cooling code. NGST will thus enable us to probe the initial mass function for star formation well below the main sequence boundary. The presence of clouds in these atmospheres can modify the spectrum in the 5 m window (and in other near-ir windows), producing a more greybody spectrum depending on the particle size. The abundant water molecule is expected to be 6

7 the primary cloud-forming constituent for T ef f < 400K; ammonia also becomes important for T ef f < 150K. At T ef f > 400K, EGP atmospheres should be relatively clear of major cloud decks, until 1000K when silicates become relevant. Whether or not clouds will play a major role below 400K is uncertain, as Jupiter shows strong 5 m emission through its very patchy water and ammonia clouds. The proposed R = 100 spectroscopic observations would provide the critical foundation for future theoretical studies which can help resolve cloud formation issues. We estimate that we will require integration times of about 10 hours per object to obtain spectra with sucient signal-to-noise to detect the major absorption bands due to water, methane, ammonia and phosphine expected in the (1, 20) m region. For the 20 brightest planetary companions, we will carry out observations with additional lters at 1.6, 4.5, 5.1, and 12.8 m, with integration times of 3 hr/lter. This program will thus require a total of 31 days of integration time. Acknowledgements We thank Mark Marley (NMSU) for computing spectra in support of this study. References Angel, R. and Burge J. 1999, NASA Napa Workshop, March 1999 Allard, F., Hauschildt, P.H., Alexander, D.R., Starreld, S. ARAA 1997, 35, 137 Burrows, A., Marley, M., Hubbard, W.B., Lunine, J.I., Guillot, T., Saumon, D., Freedman, R., Sudarsky, D., Sharp, C. 1997, ApJ, 491, 856 Cochran, W.D., Hatzes, A.P., Butler, R.P., Marcy, G. 1997, ApJ, 483, 457 Marcy, G.W., & Butler, R.P., 1998, ARAA, 36, 57 Butler, R.P., Marcy, G.W., Fischer, D.A., Brown, T.W., Contos, A.R., Korzennik, S.G., Nisenson, P., Noyes, R.W. 1999, submitted to ApJ Mayor, M. and Queloz, D. 1995, Nature, 378, 355 Nakajima, T., Oppenheimer, B.R., Kulkarni, S.R., Golimowski, D.A., Matthews, K., Durrance, S.T. 1995, Nature, 378, 463 Oppenheimer, B., Kulkarni, S.R., Matthews, K., van Kerkwijk, M.H. 1998, ApJ, 502, 932 Ortiz, J.L., Orton, G.S., Friedson, A.J., Stewart, S.T., Fisher, B.M., Spenceer, J.R. 1998, JGR, 103, No. E10, Trauger, J.T. et al. 1998, BAAS, 30,

8 Lalande pc Jupiter (M-band, 0.5 hr) 5 AU Figure 1: Simulated NGST detection of a planet of Jupiter's brightness at 5 m orbiting Lalande 21185, a nearby (2.5 pc) main-sequence M2 dwarf. The planet is assumed to have an orbital radius of 5 AU. A coronagraph is used to attenuate the glare of the primary star. The simulation uses the most up-to-date descriptions of NGST background and detector noise contributions, a 10% lter bandwidth, and 50% overall thruput. 8

9 NGST Uniqueness/Relationship to Other Facilities The substellar companions targeted in this proposal are fainter than their primary star by factors of more than 10 6, and are separated by at most a few arcseconds. To image with such high contrasts and small separations requires a large, diraction limited telescope with a very stable PSF. While 8 meter class ground-based telescopes with adaptive optics will achieve the necessary angular resolution, their Strehl and hence achievable contrast at small separations is seriously limited by uncorrected high-order phase errors that lie beyond the capabilities of practical groundbased (kilohertz rate) adaptive optical systems. Of still greater importance is the large (factors of 100 or more) reduction in sky and telescope background which NGST provides at the advantageous observing wavelength of 5 m, where the contrast between the planet and star is most favorable. While ground-based observations should progress in the coming years with detections of additional luminous brown dwarfs companions at relatively large angular separations from their parent stars, only NGST oers the prospect for directly detecting objects of Jupiter's luminosity and orbital separation in the solar neighborhood. Observing Strategy The two discovery surveys will be conducted in a 10% bandwidth lter centered at 4.8m, similar to the standard M band. The M-band is the window of choice for the survey observations because it provides the largest signal-to-noise ratio for detection. At shorter wavelengths the planet ux is signicantly lower, whereas at longer wavelengths the thermal background is signcantly higher. We have veried this conclusion by computing the coronagraphic performance at dierent wavelengths. Each target will be observed at two epochs separated by 6 months. Multiple spacecraft roll angles will be used to help discriminate detected objects from residual speckles in the NGST diraction pattern. True companions will be identied by observing their common proper motion with the central star over the 6 month baseline. In estimating our sensitivity, we have made the conservative assumption that the substellar companion is located 45 degrees of orbital longitude from maximum elongation. This makes the observation more dicult, but ensures that planets will be detected in more than 70% of the systems where they are present. Filters specic to the features in brown dwarf/jovian planet atmospheres will be needed for the spectral characterization of faint substellar companions; these will include a 1.6 m \methane-free" lter and special lters adjacent to M band to measure the spectral slope of the superthermal emission. Special Requirements Observations at multiple roll angles and at least two epochs are required for the two discovery surveys. A xed slit orientation is likely for the coronagraph's low resolution spectrograph mode. This will result in narrow observing windows during which a substellar companion could be placed on the slit while the primary star is simultaneously behind the coronagraphic mask. 9